Question

MCAT Tutorial: Human Subjects Research

Answer 1

\({\bf{Experimental~Approach:}}\)
- randomization: using an algorithm or other method to ensure that subjects are randomly assigned to treatment groups
- blinding: ensuring that the subjects and/or experimenters (single blind vs double blind) do not know which individuals are in which group
- regression analysis: fitting data to the closest mathematical model for the data
- types of variables: binary (yes or no) continuous (a non-discrete quantity) or categorical (qualitative groups)

Answer 2

\({\bf{Observational~Approach:}}\)
- cohort studies: sorting subjects based on their level of risk and assessing the differences in outcomes for those groups
ex. separating smokers and nonsmokers and measuring their rates of lung cancer
- cross-sectional studies: grouping a large sample size into two groups and assessing the differences in some metric at a single point in time
- case-control studies: separating individuals based on a desired metric and assessing those groups for differences in a particular risk factor

\({\bf{Hill's~Criteria:}}\) used to help determine whether a correlation indicates causation

- temporality: the change in the independent variable must happen before the dependent
- strength: the more variability in the dependent variable that can be explained by the independent variable, the more likely it is to be a causal relationship
- dose-relationship: the extent to which an increase in the change of the independent variable changes the dependent variable
- consistency: the prevalence of the relationship across multiply studies
- plausibility: the more reasonable the proposed relationship is, the more likely the relationship is causal. this can be partly determined by using existing research as supporting evidnece
- consideration of alternative explanations: by eliminating or discounting other explanations, the more likely the proposed causality is to be true
- experimentation: if the relationship still exists in a controlled experiment rather than an observational setting, it is more likely to be causal (this is not always possible)
- specificity: the degree to which the change in the dependent variable only results from changing the independent variable and from no other changes
- coherence: the extent to which the proposed causality is consistent with existing research

Answer 3

Anyway, that's the end of my tutorial, I hope it was a helpful resource. Source material is the Third Edition Kaplan Physics & Mathematics Prep Book for the new MCAT